组合式非周期缺陷接地结构的RBF神经网络模型

RBF Neural Network Model for Combinatorial Nonperiodic Defected Ground Structures

组合式非周期缺陷接地结构(CNPDGS)是由光子带隙结构(PBG)发展而来的,它具有结构简单、电路尺寸小、插入损耗小、设计参数少等优点。本文采用RBF神经网络建立了CNPDGS的神经网络模型。神经网络训练成功后,在其学习范围内,该模型能立刻给出任意尺寸结构的准确可靠的传输系数(S21)。结果证明神经网络建模的方法具有快速、准确、可靠等优点,具有很高的实用价值。

Combinative Nonperiodic Defected ground structures （CNPDGS） are expanded from the photonics bandgap （PBG） structures. It features simple struture, small circuit sizes, small insertion loss and less design parameters. In this paper, Radial Basis Function （RBF） artificial neural network （ANN） of CNPDGS is developed. Within the range of training, the transmission coefficient of CNPDGS at any arbitrary sizes can be obtained quickly and correctly from the ANN model that has been trained successfully. The result indicated that the way of modeling with ANN has the advantages of saving time, accuracy and reliability, and it is very useful in practice.